DARPA NVIDIA to develop a new generation of processors
Today there are many everyday objects and devices which, in its interior, which contain a computer, rather than general purpose (such as personal computers) are focused onto a discrete set of tasks and functions. Such systems are known as embedded systems and are in cars, industrial machinery or, for example, in the fleet of drones from the U.S. Army. With the idea of improving the performance of embedded systems and increase the processing power while decreasing consumption needs, DARPA has signed a contract with NVIDIA to develop a new generation of processors that substantially improves the ability to capture and process drones data from the system itself embedded in the plane.
The contract has a time horizon of five years and, during this time, nvidia developed the PERFECT program of DARPA (Power Efficiency Revolution for Embedded Computing Technologies), which will focus on developing more efficient processors 75 times (from an energy standpoint ) than currently used and, of course, improve the processing capacity. Less use, in general, can be translated into greater autonomy and whether we add more processing capabilities, the drones could analyze ad-hoc data captured by their vision sensors and provide information of much greater value to operators and troops, thus far these aircraft-autonomous robot.
As a goal, DARPA has set a rate of 75 GFLOPs processing per watt of power consumed (currently the rate is 1 per watt GFLOP) and improve manufacturing processes to further reduce the size of transistors to processors arrive at 7 nanometers (when processors now working with transistors and 28 nm).
The technologies developed under this program will transform the capabilities of embedded systems, autonomous vehicles doing much more practical and intelligent. In addition, this research will contribute to further advancement NVIDIA mobile computing applications both for government and the end consumer oriented
The project, from a technological point of view, is a major challenge for NVIDIA because it involves substantially improve processor design and manufacturing processes, tasks that will be supported by the expertise of the Universities of Utah and Virginia ( research partners), a project to reach fruition mark a turning point in the capabilities of systems that can be found in cars, trains or systems as complex as an aircraft or a drone.Tags: DARPA, Microprocessors, nvidia